Wednesday, April 15, 2009

ETF Warnings

My previous post outlined two reasons why ETFs might be more attractive than traditional low-cost indexed mutual funds.  Here are some down-side risks to check for.

Does This ETF Suck?

This isn't a risk unique to ETFs - there are lots and lots of lousy mutual funds too.  But suffice it to say, the biggest risk with an ETF is that it isn't a very good one.  Check the expense ratio, and make sure the index makes sense for your investment goals.  There are actively managed ETFs, custom-indexed ETFs, high expense ratio ETFs - lots of weird stuff.

Brokerage Fees

ETFs are "more fair" (IMHO) than mutual funds because the cost of buying and selling the shares is out front where everyone can see it: you pay a commission to buy or sell an ETF, just like you would a stock.

But this means you have a fixed cost per transaction, and this changes things compared to a mutual fund.  For example, if your 401k normally buys a small amount of 4 mutual funds every pay period, that's 104 transactions a year.  At $12.50 a transaction, you've paid $1300 in brokerage fees.  If you contribute the IRS maximum ($16,500) that's a 7% commission!  Ouch!

Simply put, heavy dollar-cost averaging over a wide number of funds isn't going to be cost-effective with ETFs.  To use ETFs, you need to make contributions in fairly large chunks.

It should be noted that mutual fund buys are not really cost-free; the mutual funds that you can buy through your broker without a sales fee are subsidized by the mutual fund itself.  Example: I can buy any Fidelity mutual fund for free via my Fidelity account because Fidelity's mutual funds have an agreement with Fidelity's brokerage.  If I buy a Van Guard mutual fund I pay $75 - way worse than a trading commission.

Typically mutual funds that are free to buy will have a higher expense ratio - part of the expense ratio will go back to "mutual fund supermarket" brokers (like Fidelity, Charles Schwab, etc.) to offset buying costs.  There is no free lunch.  (One possible exception: Fidelity heavily subsidizes its own Spartan index funds to keep the expense ratios competitive with Van Guard.  So there may be some loss leader mutual funds.)

The bottom line is: your buying pattern may have to change to use ETFs, and there is a cost to not dollar-cost averaging, so look at the big picture and the details.

The Bid-Ask Spread

Since ETFs are traded, there is a bid-ask spread - that is, the price you sell at will always be a tiny bit lower than the price you buy at on the exchange.  That spread is the profit that market makers get for providing liquidity (that is, buying and selling whenever people want to keep the market working).

A healthy ETF should have high liquidity and thus a very narrow bid-ask spread.  For example, IVV (an iShares S&P 500 ETF) has a 1 penny bid-ask spread on an $85.51 share price - that is, the spread is really, really tiny.  If the ETF is less liquid, traded less often, has fewer shares, etc. you might see wider spreads.

If the bid-ask spread is too wide, treat this like an expense; you will lose the spread each time you buy/sell.  If an ETF were to have a 1% bid-ask spread, you'd lose 1% when you buy and sell.

(A one-time loss is not as bad as recurring fee like expense ratios, but it does warrant examination.)

Arbitrage Failure/Tracking Error

When you buy some shares of an open-ended mutual fund (most mutual funds are open-ended), the mutual fund manager goes out and buys some stock (or bonds or stuff) - in other words, crunch all you want; when you buy, the manager makes more.  To understand the problem of arbitrage failure in ETFs, we have to first ask: where to ETF shares come from.

When you buy and sell shares of an ETF, you are simply buying or selling existing shares on an exchange.  But...what if everyone wants shares of IVV (an S&P 500 index ETF) at once and there just aren't enough? What keeps the price of IVV from skyrocketing above the S&P index that it should track due to the outsized demand?

The answer is in the way ETF shares are born: in-kind creation and redemption.  ETF shares are created by someone coming to the ETF with the underlying stock or bind and saying "convert these stocks that make up the S&P into ETF shares".  The ETF takes the stock and holds onto it (that's what the ETF does) and issues brand new shares.  There are now more ETF shares in existence but the ETF now has more underlying stock to back them up.

Now when someone does an in-kind creation or redemption, he or she will make a profit or loss based on the difference between the ETF share price (as traded on the market) and the price of the underlying stocks and bonds that make up the ETF.  This is what keeps an ETF's price so close to its underlying value ("net asset value" or NAV).  

If the ETF share prices were ever selling for a lot more than the underlying stocks, some clever investor could buy the stocks and sell the ETF at the same time (making a profit).  To then recover the sold ETF shares, the investor simply converts the bought stock to ETF shares.  In other words, if the ETF trading price diverges from the NAV price, there is an arbitrage opportunity.  Arbitrage keeps ETF trading and NAV prices in sync, and that's why your ETF shares are always worth the value of the underlying assets.

Well, almost always.  During the recent market crisis, some ETF shares were trading at less than the value of their underlying assets.  How did this happen?  Vanguard explains here.

Basically if the NAV of an ETF share is below the trading price you don't want to buy (you'd be over-paying) and if the NAV of an ETF share is above the trading price you don't want to sell (because theoretically you should be able to get a better price).  In practice, you're most likely to see this with bond funds, where the market for selling the underlying assets (bonds) is (a) not very fast/transparent like a stock exchange and (b) tends to seize up periodically these days.


ETFs introduce one last potential risk: you.  If you are the kind of person who might be tempted to day-trade your index funds, if  given the tools to do so, you might not want an ETF. Mutual funds are fairly clunky to buy & sell - they're terrible for day trading and speculating. ETFs don't have this weakness - you can short them, trade them every 3 minutes, and probably even buy them on margin.  If the temptation is too much, an ETF might not be a good idea.

Monday, April 13, 2009

Two Reasons to Consider ETFs

I am assuming that you already use low-cost indexed mutual funds, and assuming retirement planning (e.g. ignoring taxes).  Here are two reasons to use low-cost index-based exchanged traded funds (ETFs).  I am not arguing in favor of high cost ETFs (they do exist!) or some of the weirder indices (people make up indices and build ETFs around them all the time) - this is just a comparison of mutual funds vs. ETFs for things like the S&P 500.

WTF is an ETF?

Here are two articles explaining (or trying to explain ETFs).  But the basic idea is this: an ETF is like a public company whose business is to simply own the stocks (or bonds) in the index it tracks.  Since the company doesn't really have a business plan other than holding these stocks or bonds, the share value of the company is pretty much equal to the assets inside it.  Thus you can trade a single "stock" (the ETF shares) on an exchange instead of trading the 500 stocks that make up the S&P.  In other words, ETFs are sort of "meta-stocks".

How is this different from a mutual fund?
  • In a mutual fund, when you buy & sell shares in the mutual fund, the mutual fund buys and sells the underlying "stuff" inside the fund to raise the cash to give you your proceeds when you sell, and then buys more stuff when you give them cash to buy.
  • In an ETF, you actually buy and sell the ETF shares itself - the stocks inside the ETF don't have to be unbundled, sold separately, then rebundled each time a share of the ETF is sold.
This structure turns out to be good in two ways...

ETFs Often Have Lower Expense Ratios

When I went looking for a broad bond-index mutual fund a few years ago, I couldn't find anything cheaper than an expense ratio of 0.45.  For a bond fund, that's not very good - if I am only going to earn 3%, I don't want 15% (!) of my profits going to the managers.  You can find bond ETFs with expense ratios a lot lower - more like 0.11%.

There are a number of reasons why ETFs might have cheaper costs:
  • The market is very competitive and very hot.
  • The structure may actually be cheaper to work with.  When ETF shares change hands, only the ETF shares get sold, not the underlying instruments.*  So it is possible that ETFs are less costly to run.
  • Mutual funds have this ugly property: if I sell all my shares and get out of Fidelity's Super-Actively-Managed-Wicked-Indexed-Cool-Kids-Fund (SAMWICK?!?) and you have shares, the fund pays the brokerage fee for the sale out of the fund.  But wait ... you own the fund and I don't.  You just paid my brokerage fees.  Sucker!
  • By comparison, if I sell my ETF shares, I pay a brokerage fee to sell (like I would a stock). I pay my own costs.
On this last point, I would say the cost structures of ETFs are more fair - but it also means that the expense ratio of an ETF doesn't tell the whole story.  (Nor does the expense ratio of a mutual fund - more on that in the next post.)

The bottom line here is that if you are shopping for cheap indices to hold, the best deals seem to be in ETFs.  Van Guard has a nice set of low-cost ETFs wrapped around a number of indices -they are liquid (mostly), well funded, and have very low expense ratios.

More Control Over Buying and Selling

The way the two-step buying process of a mutual fund (step 1 - buy the fund; step 2 - the fund buys more "stuff") is implemented in the US is really lousy: I put a buy order in during the day, then the market closes, then they recalculate the fund price, and then they buy for me at the new price.*  In other words, there is a huge delay in order execution.

That didn't bother me three years ago, but times have changed, and the market jumps up and down several percentage points in a single day.  And that's where another feature of ETFs is handy: the ability to execute trades faster and place limit orders.

Since an ETF is traded like a stock, all of the things you can do with a stock are available on the ETF - in particular, you can place a limit order, saying "I would like to buy 1000 shares of this ETF when and only when the price dips below $50 a share".  You set the limit order and walk away, knowing that if the ETF falls in value (altering the balance of your portfolio) the limit order will (hopefully) kick in and change your asset allocation back to where it should be.

(In my experience limit orders don't work as perfectly as you might hope - I don't know what kind of mechanism is behind the execution process, but exchanges aren't perfectly continuous - there will be a limit to the precision of order execution.)

Similarly, if you place a market order on an ETF, it's going to go through pretty fast.  So in volatile times, with an ETF at least you know what you're buying.

ETFs also have some potentially negative features, and they're a bit more subtle than an ETF's features.  I will cover them next.

* The mutual fund situation is worse - since the NAV is calculated after hours, you only know the NAV if you buy after hours - but if you do that, they wait an entire market cycle.  So you basically never know what you're getting unless you calculate NAV yourself right before the close of market.

Monday, April 06, 2009

Isolating the Wrist

Plenty of people have written plenty of blog entries about Ultimate Frisbee, and, being a typical perennial rec-leaguer (read: not that good) I'm not really qualified to write squat about Ultimate frisbee.

But - I did manage to "fix" my back-hand throws over the last year (more or less) and, being a big nerd with no coordination who doesn't "get" any sporting activity naturally, I can write about how I did it, because the process was deliberate, intentional, and took a while.

I think you could identify a number of important elements to throwing a frisbee - spin is always at the top of the list, as well as the relationship between the disc and its flight path, how the disc is oriented at release, etc.  In my case what I was missing most was wrist isolation.

When you throw a frisbee, the wrist has a special role: many muscle groups are providing forward velocity, which will make the disc go faster (and thus farther before it falls out of the sky).  For a short throw, the elbow and shoulder might be involved; for a longer throw the whole body might rotate.

But the wrist is the only part of the body that can make the disc spin.  Why is that?  Well, the bigger the muscle group, the slower the rotational speed it produces.  I can snap my wrist almost instantly, but I can't rotate my hips 90 degrees that fast.

Thus the key to a frisbee throw is to "save" the wrist snap until the very last instant possible.  If you snap your wrist too early one of two things happens:
  • Your wrist has rotated as far as it goes and the throw isn't over.  Or more likely...
  • Your wrist rotates more slowly than it can so as to rotate at the same speed as bigger muscle groups.
I suspect that many attempts at "big" throws go wrong because the thrower synchronizes the wrist and another muscle group.

For my forehand, snapping my wrist at the last second just sort of happens - I don't know why. But for my back-hand, one of the keys to rehabilitating it was to learn to delay the wrist snap until the last minute - it was an issue of coordination.

What finally got my head in the game was to think of the arm movement and wrist snap as two distinct steps.  The exercise I do is to first step across to 10 o clock with the arm swinging as part of the movement, and then, only once my arm actually gets to the throwing position, snap my wrist to throw, as if throwing without any arm movement at all.

(A number of websites suggest first throwing wrist-only to isolate muscle groups - I think this is a good suggestion.  The next step then is to be able to move the arm and not lose the independent wrist control.)